VH2+ Antigen-Experienced B Cells in the Cerebrospinal Fluid Are Expanded and Enriched in Pediatric Anti-NMDA Receptor Encephalitis

Pediatric and adult autoimmune encephalitis (AE) are often associated with Abs to the NR1 subunit of the N-methyl-d-aspartate (NMDA) receptor (NMDAR). Very little is known regarding the cerebrospinal fluid humoral immune profile and Ab genetics associated with pediatric anti-NMDAR-AE. Using a combination of cellular, molecular, and immunogenetics tools, we collected cerebrospinal fluid from pediatric subjects and generated 1) flow cytometry data to calculate the frequency of B cell subtypes in the cerebrospinal fluid of pediatric subjects with anti-NMDAR-AE and controls, 2) a panel of recombinant human Abs from a pediatric case of anti-NMDAR-AE that was refractory to treatment, and 3) a detailed analysis of the Ab genes that bound the NR1 subunit of the NMDAR. Ag-experienced B cells including memory cells, plasmablasts, and Ab-secreting cells were expanded in the pediatric anti-NMDAR-AE cohort, but not in the controls. These Ag-experienced B cells in the cerebrospinal fluid of a pediatric case of NMDAR-AE that was refractory to treatment had expanded use of variable H chain family 2 (VH2) genes with high somatic hypermutation that all bound to the NR1 subunit of the NMDAR. A CDR3 motif was identified in this refractory case that likely drove early stage activation and expansion of naive B cells to Ab-secreting cells, facilitating autoimmunity associated with pediatric anti-NMDAR-AE through the production of Abs that bind NR1. These features of humoral immune responses in the cerebrospinal fluid of pediatric anti-NMDAR-AE patients may be relevant for clinical diagnosis and treatment.

Neuronal binding by antibodies can be influenced by low pH stress during the isolation procedure

Low pH stress and its influence on antibody binding is a common consideration among chemists, but is only recently emerging as a consideration in Immunological studies. Antibody characterizations in Multiple Sclerosis (MS), an autoimmune disease of the Central Nervous System (CNS) has revealed that antibodies in the cerebrospinal fluid (CSF) of patients with Multiple Sclerosis bind to myelin-related and non-myelin antigen targets. Many laboratories have used molecular biology techniques to generate recombinant human antibodies (rhAbs) expressed by individual B cells from healthy donors and patients with systemic autoimmune disease to identify antigen targets. This approach has been adapted within the Neuroimmunology research community to investigate antigen targets of individual B cells in the CSF of MS patients. Our laboratory determines which antibodies to clone based on their immunogenetics and this method enriches for cloning of rhAbs that bind to neurons. However, newer technologies to assist in purification of these rhAbs from culture supernatants use an acidic elution buffer which may enhance low pH stress on the antibody structure. Our laboratory routinely uses a basic elution buffer to purify rhAbs from culture supernatants to avoid low pH stress to the antibody structure. Our goal was to investigate whether acidic elution of our rhAbs using Next Generation Chromatography would impact the rhAbs' ability to bind neurons. The limited data presented here for two neuron-binding rhAbs tested indicated that acidic elution buffers used during rhAb purification impacted the ability of rhAbs with low CDR3 charge to maintain binding to neuronal targets. Reproducibility in a larger panel of rhAbs and factors underlying these observations remain untested.

Neuron-binding antibody responses are associated with Black ethnicity in multiple sclerosis during natalizumab treatment

Multiple sclerosis is an inflammatory degenerative condition of the central nervous system that may result in debilitating disability. Several studies over the past twenty years suggest that multiple sclerosis manifests with a rapid, more disabling disease course among individuals identifying with Black or Latin American ethnicity relative to those of White ethnicity. However, very little is known about immunologic underpinnings that may contribute to this ethnicity-associated discordant clinical severity. Given the importance of B cells to multiple sclerosis pathophysiology, and prior work showing increased antibody levels in the cerebrospinal fluid of Black-identifying, compared to White-identifying multiple sclerosis patients, we conducted a cohort study to determine B cell subset dynamics according to both self-reported ethnicity and genetic ancestry over time. Further, we determined relationships between ethnicity, ancestry, and neuron-binding IgG levels. We found significant associations between Black ethnicity and elevated frequencies of class-switched B cell subsets, including memory B cells; double negative two B cells; and antibody-secreting cells. The frequencies of these subsets positively correlated with West African genetic ancestry. We also observed significant associations between Black ethnicity and increased IgG binding to neurons. Our data suggests significantly heightened T cell-dependent B cell responses exhibiting increased titres of neuron-binding antibodies among individuals with multiple sclerosis identifying with the Black African diaspora. Factors driving this immunobiology may promote the greater demyelination, central nervous system atrophy and disability more often experienced by Black-, and Latin American-identifying individuals with multiple sclerosis.

Dynamic kernel matching for non-conforming data: A case study of T cell receptor datasets

Most statistical classifiers are designed to find patterns in data where numbers fit into rows and columns, like in a spreadsheet, but many kinds of data do not conform to this structure. To uncover patterns in non-conforming data, we describe an approach for modifying established statistical classifiers to handle non-conforming data, which we call dynamic kernel matching (DKM). As examples of non-conforming data, we consider (i) a dataset of T-cell receptor (TCR) sequences labelled by disease antigen and (ii) a dataset of sequenced TCR repertoires labelled by patient cytomegalovirus (CMV) serostatus, anticipating that both datasets contain signatures for diagnosing disease. We successfully fit statistical classifiers augmented with DKM to both datasets and report the performance on holdout data using standard metrics and metrics allowing for indeterminant diagnoses. Finally, we identify the patterns used by our statistical classifiers to generate predictions and show that these patterns agree with observations from experimental studies.

CSF-Derived CD4+ T-Cell Diversity Is Reduced in Patients With Alzheimer Clinical Syndrome

BACKGROUND AND OBJECTIVES

Patients with Alzheimer dementia display evidence of amyloid-related neurodegeneration. Our focus was to determine whether such patients also display evidence of a disease-targeting adaptive immune response mediated by CD4⁺ T cells. To test this hypothesis, we evaluated the CSF immune profiles of patients with Alzheimer clinical syndrome (ACS), who display clinically defined dementia.

METHODS

Innate and adaptive immune profiles of patients with ACS were measured using multicolor flow cytometry. CSF-derived CD4⁺ and CD8⁺ T-cell receptor repertoire genetics were measured using next-generation sequencing. Brain-specific autoantibody signatures of CSF-derived antibody pools were measured using array technology or ELISA. CSF from similar-age healthy controls (HCs) was used as a comparator cohort.

RESULTS

Innate cells were expanded in the CSF of patients with ACS in comparison to HCs, and innate cell expansion increased with age in the patients with ACS, but not HCs. Despite innate cell expansion in the CSF, the frequency of total CD4⁺ T cells reduced with age in the patients with ACS. T-cell receptor repertoire genetics indicated that T-cell clonal expansion is enhanced, and diversity is reduced in the patients with ACS compared with similar-age HCs.

DISCUSSION

Examination of CSF indicates that CD4⁺ T cell-mediated adaptive immune responses are altered in patients with ACS. Understanding the underlying mechanisms affecting adaptive immunity will help move us toward the goal of slowing cognitive decline.

Outcome and Immune Correlates of a Phase II Trial of High-Dose Interleukin-2 and Stereotactic Ablative Radiotherapy for Metastatic Renal Cell Carcinoma

PURPOSE

This phase II clinical trial evaluated whether the addition of stereotactic ablative radiotherapy (SAbR), which may promote tumor antigen presentation, improves the overall response rate (ORR) to high-dose IL2 (HD IL2) in metastatic renal cell carcinoma (mRCC).

PATIENTS AND METHODS

Patients with pathologic evidence of clear cell renal cell carcinoma (RCC) and radiographic evidence of metastasis were enrolled in this single-arm trial and were treated with SAbR, followed by HD IL2. ORR was assessed based on nonirradiated metastases. Secondary endpoints included overall survival (OS), progression-free survival (PFS), toxicity, and treatment-related tumor-specific immune response. Correlative studies involved whole-exome and transcriptome sequencing, T-cell receptor sequencing, cytokine analysis, and mass cytometry on patient samples.

RESULTS

Thirty ethnically diverse mRCC patients were enrolled. A median of two metastases were treated with SAbR. Among 25 patients evaluable by RECIST v1.1, ORR was 16% with 8% complete responses. Median OS was 37 months. Treatment-related adverse events (AE) included 22 grade ≥3 events that were not dissimilar from HD IL2 alone. There were no grade 5 AEs. A correlation was observed between SAbR to lung metastases and improved PFS (P = 0.0165). Clinical benefit correlated with frameshift mutational load, mast cell tumor infiltration, decreased circulating tumor-associated T-cell clones, and T-cell clonal expansion. Higher regulatory/CD8⁺ T-cell ratios at baseline in the tumor and periphery correlated with no clinical benefit.

CONCLUSIONS

Adding SAbR did not improve the response rate to HD IL2 in patients with mRCC in this study. Tissue analyses suggest a possible correlation between frameshift mutation load as well as tumor immune infiltrates and clinical outcomes.

Reconstituting T cell receptor selection in-silico

Each T cell receptor (TCR) gene is created without regard for which substances (antigens) the receptor can recognize. T cell selection culls developing T cells when their TCRs (i) fail to recognize major histocompatibility complexes (MHCs) that act as antigen presenting platforms or (ii) recognize with high affinity self-antigens derived from healthy cells and tissue. While T cell selection has been thoroughly studied, little is known about which TCRs are retained or removed by this process. Therefore, we develop an approach using TCR gene sequencing and machine learning to identify patterns in TCR protein sequences influencing the outcome of T cell receptor selection. We verify the trained models classify TCRs from developing T cells as being before selection and TCRs from mature T cells as being after selection. Our approach may provide future avenues for studying the relationship between T cell selection and conditions like autoimmune diseases.

Controlling for Confounding Factors when Diagnosing Cancer from T Cell Receptor Repertoires

Our publications have shown that T cell receptor (TCR) repertoires sequenced from tissue biopsies can be used to distinguish tumor from control tissue with > 90% classification accuracies, suggesting that TCR repertoires can be used to diagnose cancer.

TCR repertoires sequenced from blood samples of cancer patients are now publicly available, and it is tempting to use these existing samples to develop blood-based assays for diagnosing cancer. Already, research groups including us obtained promising results. However, these results are obtained by combining cancer and control samples from mismatching studies. Without controlling for age, sex, race, and even the reduced amount of blood volume collected from cancer patients relative to controls, these exciting results could merely be an artifact of confounding factors.

In this study, we control for possible confounding factors and observe that our ability to diagnose cancer from TCR repertoires sequenced from blood does not vanish. On patient-holdout cross-validations, we achieve diagnostic accuracies of 90% for colorectal cancer (controlling for depth coverage), 84% for breast cancer (controlling for depth coverage and age), and 83% for lung cancer (controlling for age, sex, race, and depth coverage).

While it is impossible to control for all potential confounding factors like variations in geographic locations, these results are encouraging. We propose future studies that (i) collect matching controls and cases and (ii) confirm TCR repertoire sequencing provides a distinct signal for diagnosing cancer independent of other blood-based assays and therefore can improve the diagnostic performance of these other blood-based assays.

Reconstituting T Cell Receptor Selection In-Silico

Each T cell receptor (TCR) gene is created without regard for which substances (antigens) the receptor can recognize. T cell selection culls developing T cells when their TCRs (i) fail to recognize major histocompatibility complexes (MHCs) that act as antigen presenting platforms or (ii) recognize with high affinity self-antigens derived from healthy cells and tissue. Here, we show that we can accurately predict from the TCR if any T cell is pre- or post-selection (and by extension if that T cell would be culled).

First, we analyze millions of TCR genes sequenced from mouse spleen using high-throughput receptor sequencing. To get examples of TCRs pre-selection, we develop a way to translate non-productive TCR genes that otherwise cannot express protein, obtaining candidate TCR protein sequences never subjected to T cell selection. To get examples of TCRs post-selection, we translate productive TCR genes sequenced from a peripheral location such as spleen, obtaining TCR protein sequences representing T cells that survived T cell selection. Next, we train a machine learning model to find sequence level patterns that discriminate between these two categories of TCR protein sequences, achieving a balanced classification accuracy of 73% (AUC of 0.8) on de-novo TCR β-chain protein sequences. This classification accuracy may be near the achievable limit using only the β-chain representing half of the TCR heterodimer.

To verify our approach, we show our machine learning model accurately predicts that (i) productive TCRs from thymus (where T selection occurs) are predominantly pre-selection and (ii) productive TCRs from peripheral tissues like skin are post-selection.

Machine Learning on Sequential Data Using a Recurrent Weighted Average

Recurrent Neural Networks (RNN) are a type of statistical model designed to handle sequential data. The model reads a sequence one symbol at a time. Each symbol is processed based on information collected from the previous symbols. With existing RNN architectures, each symbol is processed using only information from the previous processing step. To overcome this limitation, we propose a new kind of RNN model that computes a recurrent weighted average (RWA) over every past processing step. Because the RWA can be computed as a running average, the computational overhead scales like that of any other RNN architecture. The approach essentially reformulates the attention mechanism into a stand-alone model. The performance of the RWA model is assessed on the variable copy problem, the adding problem, classification of artificial grammar, classification of sequences by length, and classification of the MNIST images (where the pixels are read sequentially one at a time). On almost every task, the RWA model is found to fit the data significantly faster than a standard LSTM model.

Cell type discovery and representation in the era of high-content single cell phenotyping

Background

A fundamental characteristic of multicellular organisms is the specialization of functional cell types through the process of differentiation. These specialized cell types not only characterize the normal functioning of different organs and tissues, they can also be used as cellular biomarkers of a variety of different disease states and therapeutic/vaccine responses. In order to serve as a reference for cell type representation, the Cell Ontology has been developed to provide a standard nomenclature of defined cell types for comparative analysis and biomarker discovery. Historically, these cell types have been defined based on unique cellular shapes and structures, anatomic locations, and marker protein expression. However, we are now experiencing a revolution in cellular characterization resulting from the application of new high-throughput, high-content cytometry and sequencing technologies. The resulting explosion in the number of distinct cell types being identified is challenging the current paradigm for cell type definition in the Cell Ontology.

Results

In this paper, we provide examples of state-of-the-art cellular biomarker characterization using high-content cytometry and single cell RNA sequencing, and present strategies for standardized cell type representations based on the data outputs from these cutting-edge technologies, including “context annotations” in the form of standardized experiment metadata about the specimen source analyzed and marker genes that serve as the most useful features in machine learning-based cell type classification models. We also propose a statistical strategy for comparing new experiment data to these standardized cell type representations.

Conclusion

The advent of high-throughput/high-content single cell technologies is leading to an explosion in the number of distinct cell types being identified. It will be critical for the bioinformatics community to develop and adopt data standard conventions that will be compatible with these new technologies and support the data representation needs of the research community. The proposals enumerated here will serve as a useful starting point to address these challenges.

iWAS--A novel approach to analyzing Next Generation Sequence data for immunology

In this communication we describe a novel way to use Next Generation Sequence from the receptors expressed on T and B cells. This informatics methodology is named iWAS, for immunonome Wide Association Study, where we use the immune receptor sequences derived from T and B cells and the features of those receptors (sequences themselves, V/J gene usage, length and character each of the CDR3 sub-regions) to define biomarkers of health and disease, as well as responses to therapies. Unlike GWAS, which do not provide immediate access to mechanism, the associations with immune receptors immediately suggest possible and plausible entrée's into disease pathogenesis and treatment.

MSPrecise: a new diagnostic test for relapsing-remitting multiple sclerosis by cerebrospinal fluid B cell sequencing (HUM3P.267)

Patients with relapsing-remitting multiple sclerosis (RRMS), or who will convert to RRMS, have cerebrospinal fluid (CSF) B cells that accumulate specific replacement mutations in their rearranged VH4 gene segments as identified by single-cell Sanger sequencing. To harness this observation in a clinical setting, we have developed a diagnostic test to extract and score this genetic information from bulk CSF lymphocytes by next-generation DNA sequencing (NGS). Our current aim is to determine the accuracy of this test on a cohort of patients with varying times of RRMS onset and disease modifying therapies, as well as on a diverse cohort of patients with other neurological diseases (OND) that generally mimic the presentation of MS. Our initial testing on 39 patients has shown an accuracy of 84% in distinguishing RRMS from OND patients. As we continue to expand the size of the cohorts being tested, we will also focus on identifying differences between sub-cohorts of patients at different stages of RRMS or with particular OND diagnoses. With a growing need for a faster and more accurate diagnosis of RRMS, MSPrecise may be a useful addition to the existing diagnostic toolkit that clinicians use to achieve this goal.

The potential involvement of plasmablast antibodies in clinically isolated syndrome and multiple sclerosis (HUM3P.268)

Plasmablasts are a transient B cell subtype that secretes large amounts of antibody and arises in response to antigen exposure and T cell help. Once plasmablasts form they die off in a few days or differentiate into a long lived plasma cells, and therefore are only present during an active immune response. In our laboratory a subset of Clinically Isolated Syndrome (CIS) patients presenting with Transverse Myelitis (TM) symptoms at high risk to develop Multiple Sclerosis (MS) also have an elevated level of plasmablasts in the peripheral blood and/or the cerebrospinal fluid. Interestingly, only those patients in our cohort who have elevated plasmablast levels continue to accumulate brain MRI activity. The antibody genetics of plasmablasts from CIS TM patients are altered in comparison to those of plasmablasts responding to influenza from otherwise healthy individuals. There is increased evidence of receptor editing in CIS patients, as well as enrichment of the Antibody Gene Signature (AGS), which predicts whether CIS patients will convert to MS later in life. Antibodies from CIS plasmablasts that are enriched in overall mutations and AGS mutations recognize mouse brain by ELISA and histology, binding to multiple different cell types. By flow cytometry we have shown that many of these antibodies recognize mouse astrocyte and neuron cell lines. If these antibodies are able to bind brain tissue in vivo, they could exacerbate brain damage and contribute to disease in CIS patients.

VDJServer: a web-accessible analysis portal for immune repertoire sequence data (HUM1P.317)

Despite the importance and widespread application of detailed antigen receptor repertoire profiling via high-throughput sequencing, there is currently no suite of software tools for seamless, reproducible analysis of repertoire sequence data. Tools exist for only a subset of analysis tasks. They do not function together and are difficult to reuse without modification that requires bioinformatics expertise. This leaves researchers to perform repetitive and error-prone tasks by hand, to develop internal, idiosyncratic algorithms not easily generalizable or systematically applied, and to expend significant manual labor reformatting primary and derived data for passing between tools. This approach is not only error-prone and time- and labor-intensive, but it comes at the expense of reproducibility, both within and between research groups. We have addressed this problem by developing VDJServer, a suite of repertoire analysis tools accessible via a web interface that allows users to upload a set of antigen receptor repertoire sequences and pass them through a seamless, customizable workflow that executes all steps in the analysis and generates analysis reports complete with data summary tables, statistical analyses, figures, and workflow logs. VDJServer runs on the high-performance computing resources at the Texas Advanced Computing Center, thus enabling users to process extremely large data sets. VDJServer is funded by the NIAID and is freely available.

B cells from relapsing remitting multiple sclerosis patients modulate T cell behavior in a neuroantigen-specific manner (HUM1P.322)

The role of B cells in the pathoetiology of relapsing remitting multiple sclerosis is not well understood. B cells may contribute to MS by antibody secretion, cytokine production and antigen presentation. Antigen-specific B cells are highly efficient antigen presenting cells in the context of their cognate antigens. Since antibodies from multiple sclerosis patients bind to neuroantigens we hypothesized that neuroantigen specific memory B cells promote disease in multiple sclerosis patients by serving as antigen presenting cells for auto-reactive T cells. We observed that B cells from a subset of multiple sclerosis patients induce naïve and memory T cell proliferation in the presence of recombinant neuroantigens, myelin oligodendrocyte glycoprotein and myelin basic protein, but not MOG 35-55 peptide, in B-T cell co-cultures. Similarly, in the presence of neuroantigen proteins cells from multiple sclerosis patients secreted greater levels of IFN-γ, IL-6, -17A, -17F, -21, -22, -31, -33, TNF-α and soluble CD40L than healthy donors, while levels of IL-4 and IL-25 were similar. These data support a role for human B cells as antigen presenting cells in multiple sclerosis and highlight the role of B cells in the context of autoimmune disease beyond antibody production.

Predicting conversion to multiple sclerosis using antibody genetics: comparing Sanger and next generation sequencing methods. (TECH1P.858)

We have previously identified a distinct mutation pattern in the B cell receptors of cerebrospinal fluid B cells that can predict conversion to multiple sclerosis. This antibody gene signature, developed using Sanger sequencing on singly-sorted B cells, is limited in its clinical application by the time and effort required to obtain such single-cell repertoires. In order to increase the throughput of antibody gene signature evaluation for patients, we used 454 sequencing on bulk-sorted cerebrospinal fluid B cell pellets. We focused our analysis on identifying the mutation and distribution profiles of the repertoires obtained via both sequencing methods. Here we outline the additional filtering and data processing steps used to correct for 454 platform bias and compare the 454 and Sanger sequencing-derived repertoires which include multiple sclerosis samples sequenced by both methods in parallel. Although we observe variable amounts of background mutation and uneven sequence amplification noise by 454 sequencing for our matched samples, most of the antibody gene signature scores remain above the multiple sclerosis conversion threshold established by Sanger sequencing. We expect that additional improvements directed at the sample preparation, data filtering and sequence clustering methods presented here will help further improve the reliability of next generation sequencing-based multiple sclerosis conversion diagnosis.

Comparison of antibody gene mutation patterns in first attack optic neuritis and transverse myelitis leading to Multiple Sclerosis (47.18)

Multiple Sclerosis (MS) is the leading disease of the central nervous system and a significant, costly cause of disability in young adults. Patients at risk for MS often present with a single demyelinating event, typically called a “clinically isolated syndrome” (CIS). A CIS can occur in the brain, spinal cord (Acute Partial Transverse Myelitis, APTM), or optic nerve (Optic Neuritis, ON) at similar rates and present a challenging diagnostic and therapeutic dilemma. Stratifying CIS patients most likely to develop MS is a complicated process, but desirable since treatment with the appropriate immunomodulatory agents early in the disease course can delay long-term disability. Thus, a primary focus of our laboratory has been to develop a completely novel type of biomarker that can identify CIS patients that will develop MS. This unique biomarker is based on a pattern of antibody gene mutations (i.e. antibody gene signature or “AGS”) that is exclusive to B cells from the cerebrospinal fluid and brain lesions of MS patients that initially presented with either TM or ON. In fact, the AGS is consistently elevated in patients with one attack of ON that go on to develop definite MS, and can predict conversion to MS with 91% accuracy. More recently, we are testing the hypothesis that the AGS serves as a predictor for conversion to MS in CIS patients that present with TM. Preliminary data demonstrate that the AGS is also prevalent in CIS patients presenting with TM.

Memory B cells from relapsing remitting multiple sclerosis patients elicit functional responses by CD4+ T cells in response to neuro-antigens (135.24)

Recent evidence suggests that B and T cell interactions may be paramount in relapsing remitting multiple sclerosis (RRMS) disease pathogenesis. We hypothesized that memory B cell pools from RRMS patients may specifically harbor a subset of potent neuro-antigen presenting cells that support neuro-antigen reactive T cell expansion and activation. To assess the potential influence of neuro-antigen specific memory B cells from RRMS patients on autologous CD4+ T cell activation, we compared CD80 and HLA-DR expression, IL-10 and LTα secretion, neuro-antigen binding capacity, and functional neuro-antigen presentation by memory B cells from RRMS patients and compared these parameters to naïve B cells from RRMS patients and to memory and naïve B cells from healthy donors (HDs). We identified memory B cells from a subset of RRMS patients that were functionally capable of presenting neuro-antigen to autologous T cells. Notwithstanding the fact that the phenotypic parameters that promote efficient antigen presentation were observed to be similar between RRMS and HDs memory B cells, a corresponding capability to present neuro-antigen was not observed in HD memory B cells. Our results demonstrate for the first time that the memory B cell pool in RRMS harbors neuro-antigen specific clones that can activate T cells.

Induction of labor with vaginal prostaglandin E2 pessaries

Prostaglandin E2 vaginal pessaries (3 mg) were compared with conventional amniotomy and oxytocin infusion as a method of induction of labor in 160 patients in the study group compared with 160 in the control group. Each group consisted of 100 primigravidae and 60 multigravidae. When the features of labor, delivery and fetal status were analyzed for the study and the control groups, the patients who received PGE2 pessaries had a better outcome. The difference was more significant for those patients with a low Bishop score. Compared with controls, the PGE2-treated patients had fewer cesarean sections either for failed induction or fetal distress (P less than 0.01); the incidence of infants with low Apgar score was significantly less (P less than 0.05) and there were fewer postpartum haemorrhages (P less than 0.01).